Research map / Video / Action recognition
Animal Kingdom: A Large and Diverse Dataset for Animal Behavior Understanding
Xun Long Ng, Kian Eng Ong, Qichen Zheng, Yun Ni, Si Yong Yeo, Jun Liu
Check your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Add an approximated animal-specific feature alongside the general feature (ablation) | CARe · Unseen animals action recognition | accuracy (%): 27.3 -> 39.7 | Improved |
| Add a general feature alongside the approximated specific feature (ablation) | CARe · Unseen animals action recognition | accuracy (%): 38.2 -> 39.7 | Improved |
| Use spatially-aware tensor weights instead of scalar similarity weights (ablation) | CARe · Unseen animals action recognition | accuracy (%): 37.1 -> 39.7 | Improved |
| Use X3D with EQL instead of its cross-entropy baseline | X3D · Animal Kingdom action recognition | overall mAP: 25.25 -> 30.55 | Improved |
| Use X3D with LDAM-DRW instead of its cross-entropy baseline | X3D · Animal Kingdom action recognition | overall mAP: 25.25 -> 30.54 | Improved |
| Use X3D with Focal Loss instead of its cross-entropy baseline | X3D · Animal Kingdom action recognition | overall mAP: 25.25 -> 28.85 | Improved |
| Train with focal loss rather than cross-entropy | I3D · Animal Kingdom action recognition | overall mAP: 16.48 -> 26.49 | Improved |
| Train with focal loss rather than cross-entropy | SlowFast · Animal Kingdom action recognition | overall mAP: 20.46 -> 24.74 | Improved |
| Use HRNet-DARK rather than HRNet | Animal Kingdom pose estimation, Protocol 1 · All species | PCK@0.05: 66.06 -> 66.57 | Improved |
| Evaluate pose estimation on unseen animal subclasses rather than the mixed seen-class split | HRNet-DARK · Animal Kingdom pose estimation | PCK@0.05: 66.57 -> 40.28 | Hurt |
Closest papers by meaning
- MammalNet: A Large-Scale Video Benchmark for Mammal Recognition and Behavior Understanding (CVPR 2023 · similarity 0.91)
- LoTE-Animal: A Long Time-span Dataset for Endangered Animal Behavior Understanding (ICCV 2023 · similarity 0.89)
- Animal-Bench: Benchmarking Multimodal Video Models for Animal-centric Video Understanding (NeurIPS 2024 · similarity 0.89)
- MammAlps: A Multi-view Video Behavior Monitoring Dataset of Wild Mammals in the Swiss Alps (CVPR 2025 · similarity 0.87)
- The PanAf-FGBG Dataset: Understanding the Impact of Backgrounds in Wildlife Behaviour Recognition (CVPR 2025 · similarity 0.86)
- AniMo: Species-Aware Model for Text-Driven Animal Motion Generation (CVPR 2025 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
